Causal Additive Models with Unobserved Causal Paths and Backdoor Paths

2025-02-11

AI Summary

To address the challenge of causal direction identification under unobserved confounding—where backdoor paths remain uncontrollable due to latent variables—this paper proposes a novel method jointly constraining residual independence and conditional independence. First, it constructs an identifiable regression set to characterize causal directionality; then, it designs a composite search algorithm integrating conditional independence testing with residual independence testing. Theoretically, under the causal additive noise model assumption, the method achieves asymptotic identifiability of causal directions even in the presence of latent confounders. Empirical evaluation on standard benchmark datasets demonstrates that the proposed approach significantly outperforms existing methods in causal structure recovery accuracy. Notably, it is the first method to achieve robust and reliable causal direction inference under strong latent confounding interference.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Causal additive models have been employed as tractable yet expressive frameworks for causal discovery involving hidden variables. State-of-the-art methodologies suggest that determining the causal relationship between a pair of variables is infeasible in the presence of an unobserved backdoor or an unobserved causal path. Contrary to this assumption, we theoretically show that resolving the causal direction is feasible in certain scenarios by incorporating two novel components into the theory. The first component introduces a novel characterization of regression sets within independence between regression residuals. The second component leverages conditional independence among the observed variables. We also provide a search algorithm that integrates these innovations and demonstrate its competitive performance against existing methods.

Problem

• Identifies causal direction with hidden paths
• Uses regression residuals and observed independencies
• Develops algorithm outperforming current methods

Innovation

• Introduces regression residuals independence characterization
• Leverages conditional independence among observed variables
• Develops integrated search algorithm for causal discovery